ChatGPT Intelligent UI Arrives With GPT-6 Across All Plan Tiers

Answers that come with their own interface

OpenAI's newest ChatGPT release notes change what a chatbot reply looks like. On October 7, 2026, the company began rolling out GPT-6 with a feature it calls Intelligent UI. Answers can now mix text, visuals and interactive elements. That covers side-by-side comparisons and diagrams as well as small tools that run inside the conversation, such as calculators, bill splitters and games.11 ChatGPT picks the format by itself, and OpenAI says it will still give a plain text answer when that is the better fit.111

OpenAI's longer announcement describes the feature more broadly. Responses can include buttons, forms, charts and custom tools. The company also says the update speeds up reasoning, improves web search and strengthens safety.3 Its examples are ordinary ones. A recipe could show the cooking schedule for a Sunday roast next to the instructions, and a road trip plan could put the stops on a map with notes on which detours are worth taking.14 OpenAI says GPT-6 was trained to use a library of interface components and to make design decisions, including when interactivity helps and when plain text is enough. It also says the model's design judgment still needs work.14

The main reports tell the same story. One summary says Android Authority, MacRumors, The Decoder and iClarified all published within hours of each other on October 7. Each described ChatGPT building replies from text, visuals and interactive pieces, with the format chosen automatically.10 With that much agreement on the basic facts, the open question is how the feature works in daily use, not whether it shipped.

Who gets it, and when

The rollout is staggered, and different plans get different models. Plus, Pro, Business and Enterprise accounts got GPT-6 with Intelligent UI in the Chat tab on October 7. Free and Go users were scheduled to follow on October 8.3 Paid tiers run on GPT-6 Sol and the free and low-cost tiers run on GPT-6 Luna. OpenAI says both are tuned for everyday conversation.3 Timeline trackers and rollout tables give the same dates and tier split.710

The exceptions are worth noting. The Pro reasoning option still uses GPT-6 Astra and does not support Intelligent UI. The models behind ChatGPT Work and Codex are also unchanged in this release.1 Business customers get access workspace by workspace. Their release notes say GPT-6 runs on GPT-6 Sol in Chat, and that Intelligent UI and early-start answers work at reasoning levels from Instant up to Extra High on supported clients.6

The free-tier launch matters most. OpenAI says this GPT-6 model was built for the more than 1.2 billion people who use ChatGPT every week. GPT-6 models had reached paying customers about a month earlier.14 Before October 7, GPT-6 was available only in Work, Codex and the API, not in the everyday Chat experience.9

Answering while still thinking

Intelligent UI gets most of the attention, but the change to response timing may matter more over time. ChatGPT can now begin answering while it is still reasoning.1 OpenAI says reasoning models handled harder problems but made users wait until all the thinking was done. GPT-6 instead mixes thinking with answering and builds its reply step by step from what it has worked out so far, taking into account how long the user has been waiting.14 Business documentation puts it more simply: a reply may start while ChatGPT keeps thinking, and more detail follows.6

This looks like the more important engineering change. The interactive widgets are what people will notice, but long waits have been one of the main frustrations with reasoning models. Speed is also a wider theme. On October 6, OpenAI said GPT-6 Astra and GPT-6.1 Sol should feel about 50% faster by default for subscribers, including in partner tools that use Sign in with ChatGPT.3 On October 8, an Ultrafast mode for GPT-6.1 Sol reached the Responses API.4 In ChatGPT Work and Codex, Ultrafast is limited to the $500 Pro plan and eligible Enterprise and Edu plans.2

Where this fits in the year

Intelligent UI builds on earlier work. In January, OpenAI made everyday answers more visual, with quick graphics for sports stats, unit conversions and calculations. Key people, places and products became tappable highlights that open a side panel of facts and sources.11 In March, shopping results gained richer product cards and side-by-side comparisons of price, reviews and features.11 Each update moved ChatGPT a little further from plain paragraphs. The October release lets the model decide when to build a visual or interactive answer.

The last two weeks were especially busy. On October 1, shopping gained a virtual try-on for clothes and accessories, plus Favorites and folders for saving products in the ChatGPT Library. On iOS, document scanning can now combine several pages into one PDF.12 The try-on feature uses ChatGPT Images to generate the result from a selfie, and the reference photo is stored in personalization settings.5 On October 2, the Finances feature reached Free and Go users in the U.S. on web, iOS and Android.1 Plus and Pro users in the U.S. had gained credit score tracking in September, using Experian report access and VantageScore 3.0 insights.12 One tracker also lists paid audio uploads on October 6, which support transcription, summaries and questions about recordings. The same tracker notes a test of visual ads planned for the U.S. later in October.7

The mobile app has been updated just as often. ChatGPT for iOS version 1.2026.272 shows page previews inline in responses, opens Codex task links directly and redesigns the approval picker.15 The previous build added Home Screen widgets for recent tasks from selected computers, Home Screen and Lock Screen widgets showing remaining usage and reset times, and an option to keep the keyboard open after sending a message.2 Voice also changed in September. It now supports plugins on web, iOS and Android, and work begun in a voice call can continue in text after the call ends.12 Voice usage limits were simplified at the same time, and the separate Instant, Medium and High voice intelligence levels were retired.1

The bigger strategy

These updates should be read alongside OpenAI's late-September DevDay announcements. The Decoder described them as making ChatGPT look less like a chatbot and more like an operating system. They included an open plugin system, a shared workspace called ChatGPT Space, collaborative documents called Pages, slide tools, Slack and Microsoft Teams integrations, a 32-partner enterprise marketplace and a new Pro 500 tier. Plugin Extensions let developers build interactive panels that sit next to a conversation. Proposed MCP Events support lets plugins start automations when something happens in a connected app. Space now replaces Library for eligible Pro, Business and Enterprise accounts.5

As an editorial reading, Intelligent UI is the consumer version of the same strategy. DevDay gave developers and businesses ways to put interfaces inside ChatGPT, and the October release lets the model create interfaces for anyone, free users included. In both cases the goal is to keep users inside one app for tasks that used to need a calculator, a mapping tool or a comparison-shopping site.

Where the coverage disagrees

OpenAI's release notes, its Business documentation and the third-party trackers agree closely on dates, tiers and model names.367 The weaker material comes from more promotional outlets. One roundup of the DevDay features lists headlines saying GPT-6 Astra has 10 trillion parameters, can generate 3D games from one prompt, and has solved open quantum complexity problems, all presented as "reportedly." None of these claims appear in OpenAI's release notes, and readers should not treat them as established.

There are also small differences in wording. One explainer says Ultrafast Mode is up to eight times faster. OpenAI's own documentation says only that Ultrafast speeds up token generation, and it limits availability to specific plans.2 OpenAI's more cautious version is the one to trust.

What to watch

OpenAI admits the model's design judgment still needs work.14 That will be the real test. A chart or bill splitter that appears at the right moment saves time, while an unnecessary widget gets in the way. The Pro reasoning tier also lacks Intelligent UI for now.1 For the moment, OpenAI has put the interactive features in front of the largest number of casual users, while its most demanding users are still on the older text-first format.